How to use CRISPR/Cas9 in plants - from target site selection to DNA repair.

A tool for precise, target-specific, efficient and affordable genome editing, it is a dream for many researchers, from those who do basic research to those who use it for applied research. Since 2012, we have the tool that almost fulfils such requirements; it is based on CRISPR/Cas systems. However, even CRISPR/Cas has limitations and obstacles that might surprise its users. In this review, we focus on the most frequently used variant, CRISPR/Cas9 from Streptococcus pyogenes, and highlight the key factors affecting its mutagenesis outcomes. Firstly, factors affecting the CRISPR/Cas9 activity, such as the effect of the target sequence, chromatin state or Cas9 variant, and how long it remains in place after cleavage. Secondly, factors affecting the follow-up DNA repair mechanisms include mostly the cell type and cell cycle phase, but also, for example, the type of DNA ends produced by Cas9 cleavage (blunt/staggered). Moreover, we note some differences between using CRISPR/Cas9 in plants, yeasts and animals, as knowledge from individual kingdoms is not fully transferable. Awareness of these factors can increase the likelihood of achieving the expected results of plant genome editing, for which we provide detailed guidelines.

SPT6L, a newly discovered ancestral component of the plant RNA-directed DNA methylation pathway.

RNA-directed DNA methylation (RdDM) is driven by small RNAs (sRNAs) complementary to the nascent transcript of RNA polymerase V (Pol V). sRNAs associated with ARGONAUTE (AGO) proteins are tethered to Pol V mainly by the AGO-hook domain of its subunit NRPE1. We found, by in silico analyses, that Pol V strongly colocalizes on chromatin with another AGO-hook protein, SPT6-like (SPT6L), which is a known essential transcription elongation factor of Pol II. Our phylogenetic analysis revealed that SPT6L acquired its AGO-binding capacity already in the most basal streptophyte algae, even before the emergence of Pol V, suggesting that SPT6L might be a driving force behind the RdDM evolution. Since its emergence, SPT6L with the AGO-hook represents the only conserved SPT6 homolog in Viridiplantae, implying that the same protein is involved in both Pol II and Pol V complexes. To better understand the role of SPT6L in the Pol V complex, we characterized genomic loci where these two colocalize and uncovered that DNA methylation there is more dynamic, driven by higher levels of sRNAs often from non-canonical RdDM pathways and more dependent on chromatin modifying and remodeling proteins like MORC. Pol V loci with SPT6L are highly depleted in helitrons but enriched in gene promoters for which locally and temporally precise methylation is necessary. In view of these results, we discuss potential roles of multiple AGO-hook domains present in the Pol V complex and speculate that SPT6L mediates de novo methylation of naïve loci by interconnecting Pol II and Pol V activities.

Integrated In Situ Fabrication of CuO Nanorod-Decorated Polymer Membranes for the Catalytic Flow-Through Reduction of p-Nitrophenol.

We developed a novel method to fabricate copper nanorods in situ in a poly(ether sulfone) (15 wt %) casting solution by a sonochemical reduction of Cu2+ ions with NaBH4. The main twist is the addition of ethanol to remove excess NaBH4 through Cu(0) catalyzed ethanolysis. This enabled the direct use of the resulting copper-containing casting dispersions for membrane preparation by liquid nonsolvent-induced phase separation and led to full utilization of the copper source, generating zero metal waste. We characterized the copper nanorods as presented in the membranes via scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDX), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and UV/vis spectroscopy. We could demonstrate that the rapid immobilization from reducing conditions led to the membrane incorporation of copper nanorods in a state of high reactivity, which also promoted the complete oxidation to CuO after fabrication. We further observed a large aspect ratio and crystal straining of the nanorods, likely resulting from growth around the matrix polymer. The entanglement with poly(ether sulfone) further facilitated a selective presentation at the pore surface of the final CuO-decorated membranes. The membranes also exhibit high water permeances of up to 2800 L/m2hbar. Our catalytic membranes achieved exceptionally high activities in the aqueous flow-through reduction of p-nitrophenol (p-NP), with turnover frequencies of up to 115 h-1, even surpassing those of other state-of-the-art catalytic membranes that incorporate Pd or Ag. Additionally, we demonstrated that catalytic hydrolysis of the reducing agent in water can lead to hydrogen gas formation and blocking of active sites during continuous catalytic p-NP hydrogenation. We illustrated that the accompanying conversion loss can be mitigated by facilitated gas transport in the water-filled pores, which is dependent on the orientation of the pore size gradient and the flow direction.

[Xe(OTeF5)(pyF)]+: a strong oxidizing xenonium(II) teflate cation with N-donor bases.

Herein we report on the formation of the adduct salts [Xe(OTeF5)(pyF)][Al(OTeF5)4] (pyF = C5F5N, C5H3F2N) by abstraction of an -OTeF5 group from Xe(OTeF5)2 with the Lewis superacid Al(OTeF5)3 and subsequent adduct formation of the generated [XeOTeF5]+ cation with fluorinated pyridines. These salts represent the first xenonium cations with the weakly coordinating [Al(OTeF5)4]- anion. The strong oxidizing property of these compounds is further assessed.

Facile Synthesis of Catechol-Containing Polyacrylamide Copolymers: Synergistic Effects of Amine, Amide and Catechol Residues in Mussel-Inspired Adhesives.

The straightforward synthesis of polyamide-derived statistical copolymers with catechol, amine, amide and hydroxy residues via free radical polymerization is presented. In particular, catechol, amine and amide residues are present in natural mussel foot proteins, enabling strong underwater adhesion due to synergistic effects where cationic residues displace hydration and ion layers, followed by strong short-rang hydrogen bonding between the catechol or primary amides and SiO2 surfaces. The present study is aimed at investigating whether such synergistic effects also exist for statistical copolymer systems that lack the sequence-defined positioning of functional groups in mussel foot proteins. A series of copolymers is established and the adsorption in saline solutions on SiO2 is determined by quartz crystal microbalance measurements and ellipsometry. These studies confirm a synergy between cationic amine groups with catechol units and primary amide groups via an increased adsorptivity and increased polymer layer thicknesses. Therefore, the free radical polymerization of catechol, amine and amide monomers as shown here may lead to simplified mussel-inspired adhesives that can be prepared with the readily scalable methods required for large-scale applications.

Role of sesquiterpenes in biogenic new particle formation.

Biogenic vapors form new particles in the atmosphere, affecting global climate. The contributions of monoterpenes and isoprene to new particle formation (NPF) have been extensively studied. However, sesquiterpenes have received little attention despite a potentially important role due to their high molecular weight. Via chamber experiments performed under atmospheric conditions, we report biogenic NPF resulting from the oxidation of pure mixtures of ß-caryophyllene, α-pinene, and isoprene, which produces oxygenated compounds over a wide range of volatilities. We find that a class of vapors termed ultralow-volatility organic compounds (ULVOCs) are highly efficient nucleators and quantitatively determine NPF efficiency. When compared with a mixture of isoprene and monoterpene alone, adding only 2% sesquiterpene increases the ULVOC yield and doubles the formation rate. Thus, sesquiterpene emissions need to be included in assessments of global aerosol concentrations in pristine climates where biogenic NPF is expected to be a major source of cloud condensation nuclei.

An annotated human blastocyst dataset to benchmark deep learning architectures for in vitro fertilization.

Medical Assisted Reproduction proved its efficacy to treat the vast majority forms of infertility. One of the key procedures in this treatment is the selection and transfer of the embryo with the highest developmental potential. To assess this potential, clinical embryologists routinely work with static images (morphological assessment) or short video sequences (time-lapse annotation). Recently, Artificial Intelligence models were utilized to support the embryo selection procedure. Even though they have proven their great potential in different in vitro fertilization settings, there is still considerable room for improvement. To support the advancement of algorithms in this research field, we built a dataset consisting of static blastocyst images and additional annotations. As such, Gardner criteria annotations, depicting a morphological blastocyst rating scheme, and collected clinical parameters are provided. The presented dataset is intended to be used to train deep learning models on static morphological images to predict Gardner's criteria and clinical outcomes such as live birth. A benchmark of human expert's performance in annotating Gardner criteria is provided.

Sulfadiazine and phosphinothricin selection systems optimised for the transformation of tobacco BY-2 cells.

KEY MESSAGE: We extended the applicability of the BY-2 cell line as a model by introducing two new selection systems. Our protocol provides guidelines for optimising Basta selection in other recalcitrant models. Tobacco BY-2 cell line is the most commonly used cytological model in plant research. It is uniform, can be simply treated by chemicals, synchronised and easily transformed. However, only a few selection systems are available that complicate advanced studies using multiple stacked transgenes and extensive gene editing. In our work, we adopted for BY-2 cell line two other selection systems: sulfadiazine and phosphinothricin (PPT, an active ingredient of Basta herbicide). We show that sulfadiazine can be used in a wide range of concentrations. It is suitable for co-transformation and subsequent double selection with kanamycin or hygromycin, which are standardly used for BY-2 transformation. We also have domesticated the sulfadiazine resistance for the user-friendly GoldenBraid cloning system. Compared to sulfadiazine, establishing selection on phosphinothricin was considerably more challenging. It did not work in any concentration of PPT with standardly cultured cells. Since the selection is based on blocking glutamine synthetase and consequent ammonium toxicity and deficiency of assimilated nitrogen, we tried to manipulate nitrogen availability. We found that the PPT selection reliably works only with nitrogen-starved cells with reduced nitrate reserves that are selected on a medium without ammonium nitrate. Both these adjustments prevent the release of large amounts of ammonium, which can toxify the entire culture in the case of standardly cultured cells. Since high nitrogen reserves can be a common feature of in vitro cultures grown on MS media, nitrogen starvation could be a key step in establishing phosphinothricin resistance in other plant models.

Highly synchronized cortical circuit dynamics mediate spontaneous pain in mice.

Cortical neural dynamics mediate information processing for the cerebral cortex, which is implicated in fundamental biological processes such as vision and olfaction, in addition to neurological and psychiatric diseases. Spontaneous pain is a key feature of human neuropathic pain. Whether spontaneous pain pushes the cortical network into an aberrant state and, if so, whether it can be brought back to a "normal" operating range to ameliorate pain are unknown. Using a clinically relevant mouse model of neuropathic pain with spontaneous pain-like behavior, we report that orofacial spontaneous pain activated a specific area within the primary somatosensory cortex (S1), displaying synchronized neural dynamics revealed by intravital two-photon calcium imaging. This synchronization was underpinned by local GABAergic interneuron hypoactivity. Pain-induced cortical synchronization could be attenuated by manipulating local S1 networks or clinically effective pain therapies. Specifically, both chemogenetic inhibition of pain-related c-Fos-expressing neurons and selective activation of GABAergic interneurons significantly attenuated S1 synchronization. Clinically effective pain therapies including carbamazepine and nerve root decompression could also dampen S1 synchronization. More important, restoring a "normal" range of neural dynamics through attenuation of pain-induced S1 synchronization alleviated pain-like behavior. These results suggest that spontaneous pain pushed the S1 regional network into a synchronized state, whereas reversal of this synchronization alleviated pain.

Dendritic Mechanisms for In Vivo Neural Computations and Behavior.

Dendrites receive the vast majority of a single neuron's inputs, and coordinate the transformation of these signals into neuronal output. Ex vivo and theoretical evidence has shown that dendrites possess powerful processing capabilities, yet little is known about how these mechanisms are engaged in the intact brain or how they influence circuit dynamics. New experimental and computational technologies have led to a surge in interest to unravel and harness their computational potential. This review highlights recent and emerging work that combines established and cutting-edge technologies to identify the role of dendrites in brain function. We discuss active dendritic mediation of sensory perception and learning in neocortical and hippocampal pyramidal neurons. Complementing these physiological findings, we present theoretical work that provides new insights into the underlying computations of single neurons and networks by using biologically plausible implementations of dendritic processes. Finally, we present a novel brain-computer interface task, which assays somatodendritic coupling to study the mechanisms of biological credit assignment. Together, these findings present exciting progress in understanding how dendrites are critical for in vivo learning and behavior, and highlight how subcellular processes can contribute to our understanding of both biological and artificial neural computation.

Dataset providing boundary conditions of an experimental test bench to numerically investigate flame wall interactions using CFD.

The data provided allows other research groups to numerically investigate the flame wall interactions of an experimental test bench at the University of the Bundeswehr Munich. Numerical simulations can then be compared to the experimental results to test new models. Three data sets are available. The first data set contains the inflow boundary conditions created by the burner into the test section. The fields of interest are the velocity and RMS-velocity fields. Large-Eddy-Simulation (LES) using ANSYS Fluent was used to collect this data. The second data set provides the inflow boundary condition from the fuel injector into the test section using methane at a momentum ratio of I = 10. LES using OpenFOAM was used to create this data. The third data set provides the temperature distribution at the wall of the test section when injecting methane at a momentum ratio of I = 10. The temperature is provided along the wall ranging from +/- 25 mm in the lateral and 0 - 220 mm in the axial direction. The data was derived from wall-embedded thermocouples by applying the inverse heat conduction method using MATLAB and COMSOL.

Variations in the thermal conductivity of magnetosensitive elastomers by magnetically induced internal restructuring.

Magnetosensitive elastomers respond to external magnetic fields by changing their stiffness and shape. These effects result from interactions among magnetized inclusions that are embedded within an elastic matrix. Strong external magnetic fields induce internal restructuring, for example the formation of chain-like aggregates. However, such reconfigurations affect not only the overall mechanical properties of the elastomers but also the transport through such systems. We concentrate here on the transport of heat, that is thermal conductivity. For flat, thin model systems representing thin films or membranes and modeled by bead-spring discretizations, we evaluate the internal restructuring in response to magnetization of the particles. For each resulting configuration, we evaluate the associated thermal conductivity. We analyze the changes in heat transport as a function of the strength of magnetization, particle number, density of magnetizable particles (at fixed overall particle number), and aspect ratio of the system. We observe that varying any one of these parameters can induce pronounced changes in the bulk thermal conductivity. Our results motivate future experimental and theoretical studies of systems with magnetically tunable thermal but also electric conductivity-both of which have only rarely been addressed so far.

DNA methylation can alter CRISPR/Cas9 editing frequency and DNA repair outcome in a target-specific manner.

The impact of epigenetic modifications on the efficacy of CRISPR/Cas9-mediated double-stranded DNA breaks and subsequent DNA repair is poorly understood, especially in plants. In this study, we investigated the effect of the level of cytosine methylation on the outcome of CRISPR/Cas9-induced mutations at multiple Cas9 target sites in Nicotiana benthamiana leaf cells using next-generation sequencing. We found that high levels of promoter methylation, but not gene-body methylation, decreased the frequency of Cas9-mediated mutations. DNA methylation also influenced the ratio of insertions and deletions and potentially the type of Cas9 cleavage in a target-specific manner. In addition, we detected an over-representation of deletion events governed by a single 5'-terminal nucleotide at Cas9-induced DNA breaks. Our findings suggest that DNA methylation can indirectly impair Cas9 activity and subsequent DNA repair, probably through changes in the local chromatin structure. In addition to the well described Cas9-induced blunt-end double-stranded DNA breaks, we provide evidence for Cas9-mediated staggered DNA cuts in plant cells. Both types of cut may direct microhomology-mediated DNA repair by a novel, as yet undescribed, mechanism.

One-step methylation of aromatic phosphorus heterocycles: synthesis and crystallographic characterization of a 1-methyl-phosphininium salt.

For the first time, the direct synthesis of 1-methyl-phosphininium salts has been achieved by reacting aromatic λ3,σ2-phosphinines with the readily available dimethyl chloronium salt [(CH3)2Cl]+[Al(OTeF5)4]-. The remarkably high electrophilicity of the alkylation reagent in combination with the weakly coordinating pentafluoro-orthotelluratoaluminate anion offers excellent conditions for this one-step approach. Our simple and quantitative access to 1-methyl-phosphininium salts will pave the way to explore the chemistry of such reactive species in more detail.

High-Strength Suture Tapes Are Biomechanically Stronger Than High-Strength Sutures Used in Rotator Cuff Repair.

PURPOSE: To assess the mechanical properties, tendon resistance to suture cutout, and knot size of a broad variety of high-strength sutures and tapes available for arthroscopic rotator cuff repair. METHODS: Nine different types of high-strength sutures and tapes for arthroscopic rotator cuff repair were studied: 6 were tapes (FiberTape, Hi-Fi Tape, Permatape, SutureTape, UltraTape, and XBraid TT), and 3 were sutures (Dynacord, FiberWire, and Ultrabraid). First, mechanical tensile testing of suture loops (n = 6) was performed. Second, the suture material was passed through an intact human cadaveric rotator cuff tendon (supraspinatus or infraspinatus), and cyclic as well as load-to-failure testing was performed, 8 times for each suture or tape. Statistical analysis of groups (tapes vs sutures) and between each suture and each tape was performed. RESULTS: Material testing revealed significant differences with superior mechanical properties of tapes compared with sutures regarding load for 3 mm of displacement (201 N vs 84 N, P < .0001), displacement at 200 N (3.6 mm vs 6.6 mm, P < .0001), stiffness (46 N/mm vs 25 N/mm, P < .0001), and ultimate load to failure (509 N vs 288 N, P < .0001). FiberTape showed the highest ultimate load to failure (805.5 ± 36.1 N), the highest load necessary for 3 mm of displacement (376.2 ± 19.1 N), and the lowest displacement at a 200-N load (2 ± 0.1 mm). Permatape had the highest stiffness (58.5 ± 5.3 N/mm). FiberTape had the highest knot height (9.5 ± 1.3 mm) and knot width (3.8 ± 0.7 mm) of a suture loop with 7 square knots. The typical failure mode in the cadaveric study part was tendon cut through. CONCLUSIONS: Biomechanical in vitro testing showed that high-strength suture tapes compared with regular high-strength sutures have significantly better mechanical properties in both dry-laboratory testing and human cadaveric rotator cuff tendon pullout testing. FiberTape outperformed the other tapes and the sutures used in this analysis. Nonetheless, differences in tendon testing appeared to be less substantial than in dry-laboratory testing, and FiberTape had the highest knot height and width. CLINICAL RELEVANCE: FiberTape revealed the most favorable biomechanical performance in dry-laboratory and rotator cuff tendon testing. It may provide the best repair strength in vivo; however, it also has the largest knot size.

Is There an Association Between Bone Microarchitecture and Fracture in Patients who were Treated for High-grade Osteosarcoma? A Controlled Study at Long-term Follow-up Using High-resolution Peripheral Quantitative CT.

BACKGROUND: Neoadjuvant chemotherapy in patients with primary osteosarcoma improves survival rates, but it also causes side effects in various organs including bone. Low bone mineral density (BMD) can occur owing partly to chemotherapy or limited mobility. This can cause a higher risk of fractures compared with those who do not receive such treatment. Changes in BMD alone cannot explain the propensity of fractures. Studying microarchitectural changes of bone might help to understand the effect. QUESTIONS/PURPOSES: (1) Do patients who were treated for osteosarcoma (more than 20 years previously) have low BMD? (2) Do these patients experience more fractures than controls who do not have osteosarcoma? (3) What differences in bone microarchitecture are present between patients treated for high-grade osteosarcoma and individuals who have never had osteosarcoma? METHODS: We contacted 48 patients who were treated for osteosarcoma and who participated in an earlier study. These patients underwent multimodal treatment including chemotherapy more than 20 years ago. Of the original patient group, 60% (29 of 48) were missing, leaving 40% (19 of 48) available for inclusion in this study; all 19 agreed to participate. There were nine men and 10 women with a mean age of 46 ± 4 years and a mean time from surgery to examination of 28 ± 3 years. BMD was measured by dual-energy x-ray absorptiometry, and any fracture history was assessed using a questionnaire. Additionally, high-resolution peripheral quantitative CT was performed to compare the groups in terms of microarchitectural changes, such as cortical and trabecular area, cortical and trabecular thickness, cortical porosity, and endocortical perimeter. Participants in the control group were selected from a cohort consisting of a population-based random sample of 499 healthy adult women and men. Osteoporosis or low BMD was not an exclusion criterion for entering this study; however, the patients in the control group were selected based on a normal BMD (that is, T score > -1.0 at both the spine and hip). Also, the participants were matched based on age and sex. Differences between patients and controls were assessed using the Wilcoxon rank sum test for continuous variables and a chi-square test for categorical variables. A multiple regression analysis was performed. Model assumptions were checked using histograms and quantile-quantile plots of residuals. RESULTS: Twelve of 19 patients who were treated for osteosarcoma had either osteopenia (eight patients) or osteoporosis (four patients). More patients with osteosarcoma reported sustaining fractures (11 of 19 patients) than did control patients (2 of 19 controls; p < 0.001). Among all microarchitectural parameters, only the endocortical perimeter was increased in patients compared with the control group (75 ± 15 mm versus 62 ± 18 mm; p = 0.04); we found no differences between the groups in terms of cortical and trabecular area, cortical and trabecular thickness, or cortical porosity. CONCLUSION: Although patients who were treated for osteosarcoma had osteopenic or osteoporotic BMD and a higher proportion of patients experienced fractures than did patients in the control group, we could not confirm differences in microarchitectural parameters using high-resolution peripheral quantitative CT. Therefore, it seems that bone geometry and microstructural parameters are not likely the cause of the increased proportion of fractures observed in our patients who were treated for osteosarcoma. Until we learn more about the bone changes associated with chemotherapy in patients with osteosarcoma, we recommend that patients undergo regular BMD testing, and we recommend that physicians consider osteoporosis treatment in patients with low BMD. These data might provide the impetus for future multicenter prospective studies examining the association between chemotherapy and bone microarchitecture. LEVEL OF EVIDENCE: Level III, therapeutic study.

High Cortico-Trabecular Transitional Zone Porosity and Reduced Trabecular Density in Men and Women with Stress Fractures.

To determine whether stress fractures are associated with bone microstructural deterioration we quantified distal radial and the unfractured distal tibia using high resolution peripheral quantitative computed tomography in 26 cases with lower limb stress fractures (15 males, 11 females; mean age 37.1 ± 3.1 years) and 62 age-matched healthy controls (24 males, 38 females; mean age 35.0 ± 1.6 years). Relative to controls, in men, at the distal radius, cases had smaller cortical cross sectional area (CSA) (p = 0.012), higher porosity of the outer transitional zone (OTZ) (p = 0.006), inner transitional zone (ITZ) (p = 0.043) and the compact-appearing cortex (CC) (p = 0.023) while trabecular vBMD was lower (p = 0.002). At the distal tibia, cases also had a smaller cortical CSA (p = 0.008). Cortical porosity was not higher, but trabecular vBMD was lower (p = 0.001). Relative to controls, in women, cases had higher distal radial porosity of the OTZ (p = 0.028), ITZ (p = 0.030) not CC (p = 0.054). Trabecular vBMD was lower (p = 0.041). Distal tibial porosity was higher in the OTZ (p = 0.035), ITZ (p = 0.009), not CC. Stress fractures are associated with compromised cortical and trabecular microstructure.

A Swept source optical coherence tomography angiography study: Imaging artifacts and comparison of non-perfusion areas with fluorescein angiography in diabetic macular edema.

PURPOSE: Swept Source Optical coherence tomography angiography (SS-OCTA) is a novel technique to visualize perfusion and vascular changes like ischemia in patients with diabetic retinopathy. The aim of this study was to compare non-perfusion areas on conventional fluorescein angiography (FA) with those on SS-OCTA using detailed manual annotation in patients with diabetic macular edema (DME) and to evaluate possible artifacts caused by DME on SS-OCTA. METHODS: 27 eyes of 21 patients with DME were analyzed in this prospective, cross-sectional study; on all, standard ophthalmological examination, SS-OCTA and FA imaging were performed. Early-phase FA and SS-OCTA images were analyzed for capillary dropout and foveal avascular zone (FAZ) was measured on both modalities. Artifacts in SS-OCTA imaging caused by DME were marked and analyzed. RESULTS: The mean age of the patients was 62.6 ± 11.5 years. On FA the mean size of the annotated non-perfusion areas was 0.14 ± 0.31 mm2 whereas the mean size in SS-OCTA was 0.04 ± 0.13 mm2; areas marked on FA were statistically significantly larger than on SS-OCTA (p<0.01). Mean size of FAZs was similar between FA and OCTA images. (p = 0.91). Seven eyes (25.9 percent) showed imaging artifacts due to DME in SS-OCTA. CONCLUSION: SS-OCTA is a valid tool to analyze capillary perfusion status of patients with DME, although areas of non-perfusion were measured smaller than in conventional FA. More non-perfusion areas were found on SS-OCTA images. FAZ measurements were similar using the two modalities. However, SS-OCTA is prone to artifacts and therefore requires reviewing of imaging results: up to 25 percent of the analyzed eyes showed artifacts on OCTA, which occurred in the areas of diabetic macular edema and did not correspond to capillary drop out.

Evaluation of Deep Learning Architectures for Complex Immunofluorescence Nuclear Image Segmentation.

Separating and labeling each nuclear instance (instance-aware segmentation) is the key challenge in nuclear image segmentation. Deep Convolutional Neural Networks have been demonstrated to solve nuclear image segmentation tasks across different imaging modalities, but a systematic comparison on complex immunofluorescence images has not been performed. Deep learning based segmentation requires annotated datasets for training, but annotated fluorescence nuclear image datasets are rare and of limited size and complexity. In this work, we evaluate and compare the segmentation effectiveness of multiple deep learning architectures (U-Net, U-Net ResNet, Cellpose, Mask R-CNN, KG instance segmentation) and two conventional algorithms (Iterative h-min based watershed, Attributed relational graphs) on complex fluorescence nuclear images of various types. We propose and evaluate a novel strategy to create artificial images to extend the training set. Results show that instance-aware segmentation architectures and Cellpose outperform the U-Net architectures and conventional methods on complex images in terms of F1 scores, while the U-Net architectures achieve overall higher mean Dice scores. Training with additional artificially generated images improves recall and F1 scores for complex images, thereby leading to top F1 scores for three out of five sample preparation types. Mask R-CNN trained on artificial images achieves the overall highest F1 score on complex images of similar conditions to the training set images while Cellpose achieves the overall highest F1 score on complex images of new imaging conditions. We provide quantitative results demonstrating that images annotated by under-graduates are sufficient for training instance-aware segmentation architectures to efficiently segment complex fluorescence nuclear images.